[mertimiks]

So I placed a duct (diffuser) around my wind turbine, and the air flow is turbulent. So do you guys have any recommendations for which turbulence model should I used, and why?

[Far]

SST with automatic wall treatment. Keep y+ below 10

[mertimiks]

Originally Posted by Far

SST with automatic wall treatment. Keep y+ below 10

Why is that? What do you mean by y+ below 10?

[Zigainer]

Originally Posted by mertimiks

Why is that? What do you mean by y+ below 10?

Sorry, but that is real basics ... you should understand that, because otherwise you won´t know how to get a good solution.

http://www.cfd-online.com/Wiki/Dimensionless_wall_distance_%28y_plus%29

[diamondx]

This is what i learned recently, someone please corrects me if i'm wrong:
Y+ is an non-dimentional parameter that determines how refined your mesh is near a wall and therefore how accurate you can be (regarding boundary layer), the more refine the more accurate, small value means good accuracy in general below 10 and not less than 0.
There is a tool in cfd-online that allows you to determine the height of the first cell near the wall depending on the Y+ you want. just google Y+ estimator.

[Zigainer]

Originally Posted by diamondx

Y+ is an non-dimentional parameter that determines how refined your mesh is near a wall

yes

Originally Posted by diamondx

and therefore how accurate you can be (regarding boundary layer), the more refine the more accurate, small value means good accuracy in general below 10 and not less than 0.

y+ is important for wall treatment. To less cells for the boundary layer (not only y+ of the first cell, also number of cells within the boundary layer) will not resolve the boundary layer correctly and therefore your solution is not good. If you use "wall function" instead of resolve the boundary layer, your y+ can be lager (y+ ... 20-200) and you do not need so many cells within the boundary layer. low-Re mesh need y+<1 to be accurate.

Originally Posted by diamondx

and not less than 0.

y+ can´t be less than 0